Article (Published Version)

Article

On the Form and Content of the Certificates of Pagan Sacrifice

SCHUBERT, Paul

Abstract

Cloes examination of the format of certificates for pagan sacrifice found in Egypt, dating from the so-called 'persecution of Decius' of AD 250.

Reference

SCHUBERT, Paul. On the Form and Content of the Certificates of Pagan Sacrifice. Journal of Roman Studies, 2016, vol. 106, p. 172-198

DOI : 10.1017/S0075435816000617

Available at: http://archive-ouverte.unige.ch/unige:89489

Disclaimer: layout of this document may differ from the published version.

1 / 1 On the Form and Content of the Certicates of Pagan Sacrice1

PAUL SCHUBERT

ABSTRACT

Certicates of sacrice (libelli) were produced during the so-called persecution of Decius (A.D. 250), which is documented through the testimony of Christian authors and through original certicates preserved on papyrus. The aim of this article is to offer a more detailed perspective on some specic points in the procedure as regards the production of the papyri. Although Decius’ edict did not produce an instant and decisive change in the religious balance of the Empire, the procedure that was put in place nonetheless testies to the effectiveness of the existing structure, and also to the capacity of the ofcials to adapt this structure so as to obtain maximum compliance from the population.

Keywords: Decius; persecution; libelli; certicates of sacrice; papyri; Roman Egypt; Theadelphia

I INTRODUCTION

Much has already been said — notably in this journal2 — on the so-called libelli, i.e. certicates for pagan sacrice issued under the reign of Decius in the summer of A.D. 250. By an imperial edict, all individuals in the Roman Empire were required to publicly state that they had regularly sacriced to the gods in the past; they also had to perform a sacrice before their local authorities. This was then conrmed in writing by a certicate issued in the name of every individual or his household. Besides the literary evidence provided by several Christian authors, we possess forty-six such certicates preserved on papyri from Middle Egypt.3 The aim of this article is not to reconsider the whole interpretation of the edict that led to the writing of libelli, but to offer a more detailed perspective on some specic points in the procedure as regards the production of the certicates preserved on papyrus. In order to do this, we shall combine an examination of the form and of the contents of the documents. It will be shown that, although Decius’ edict did not produce an instant and decisive change in the religious balance of the Empire, his administration is not to be blamed for this: on the contrary, the procedure that was put into place testies not only to the effectiveness of the existing structure, but also to the capacity of the ofcials to adapt this structure

1 I wish to thank Professor Adalberto Giovannini (Geneva), who read a rst version of this article. My gratitude also extends to all four anonymous readers for their helpful comments. 2 See Rives 1999. 3 All the libelli known to date were listed and numbered in the most recent publication of such a document (4 =P. Lips. II 157). Below, we shall use the same numbering; for the convenience of the reader, a list is also provided at the end of this article.

JRS 106 (2016), pp. 172–198. © The Author(s) 2016. Published by The Society for the Promotion of Roman Studies. doi:10.1017/S0075435816000617 Downloaded from http:/www.cambridge.org/core. University of Geneva, on 21 Nov 2016 at 12:36:55, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/S0075435816000617 ON THE FORM AND CONTENT OF THE CERTIFICATES OF PAGAN SACRIFICE 173

so as to obtain maximum compliance from the population, down to the level of a village, in a coherent fashion. This will be made clear notably by focusing on two specic elements of the procedure, namely the consumption of sacriced meat and the request for a signature. The rst of these two elements must be considered in the light both of ancient tradition and of the Christian perception of ingestion as a reinterpretation of this tradition. Although the perspective chosen here is quite different from that followed by James Rives, the ndings do not necessarily contradict his conclusions. Rather, they allow us to better understand how, at the level of individuals in their villages, the edict was implemented by the administration. Rives gives a thorough account of the event, mostly from the wide perspective of the religion of the Empire. Some key elements should be recalled here for the sake of clarity. Rives rightly emphasizes from the outset the precedence of performance over belief: what seemed to matter was, above all, the requirement that every citizen should accomplish a sacrice in a prescribed way, and that the sacrice be duly recorded. The insistence on active involvement, however, ‘ran counter to traditional religious organization’.4 Neither the testimony of Christian authors nor the papyri have preserved the wording of Decius’ edict in any detail.5 A comparison of the certicates preserved on papyri should, however, allow us to identify some key elements that were no doubt included in the text of the edict. Although our Christian sources mention harsh punishment inicted on those who did not comply with the orders, we cannot tell for sure if repression was explicitly mentioned in the edict. It may have been caused by over-zealous ofcials who acted beyond the call of duty. Certicates were issued not only in Egypt, but also in North Africa and in Rome, which makes it likely that the procedure was in fact applied throughout the Empire.6 Graeme Clarke states that ‘in making his sweeping edict Decius may have given little serious forethought to the practical administration of his orders in, say, the backwoods of Britannia or the wilds of Mauretania: such problems were to be left to the local administration in the provincial civitates to be dealt with in such a way as they could’.7 Examination of the procedure as it appears in the libelli preserved on papyri will conrm both the reliability of the emperor’s administration, and the degree of exibility on the part of the ofcials who had to enforce the orders. Rives states that ‘it remains very likely that [Decius’] motivations in issuing the decree were largely traditional’.8 The brevity of his reign, the rather sudden stop of the operation after the summer of 250, as well the paucity of our sources make it difcult to understand why this measure was initiated and what the general feeling in the population of the Empire could have been.9 That the perfomance of a sacrice was ordered throughout the Empire seems clear enough; but Rives rightly points out that ‘none of the evidence, not even the certicates of sacrice, provides any hint about the reason for the sacrice’.10 He nonetheless states that ‘by dening the minimal cult behaviour expected of all Romans, Decius was in effect establishing a kind of orthopraxy’.11 The papyri do not help to better understand the reasons that lay behind Decius’ edict, but they show that the emperor was able to turn this into a huge administrative operation. To that effect, he was able to rely on a tested machinery that allowed some degree of exibility at the local level, provided

4 See Rives 1999: 135 and 145. 5 See Clarke 1984: 25. There is little doubt that this was an edict; see Rives 1999: 137, n. 12. 6 See Rives 1999: 141. Clarke 1984: 28 cites evidence for an impact of the edict ‘in Spain, Gaul, Italy, Sicily, Africa, Egypt, Palestine, Syria, Pontus and almost certainly Asia’. 7 See Clarke 1984: 30. 8 See Rives 1999: 143. 9 Clarke 1969: 63, n. 1: ‘How the persecution ended is a mystery — there is no hint of a sudden, general amnesty. Cyprian’s wording suggests it merely petered out.’ 10 See Rives 1999: 147. 11 See Rives 1999: 153.

Downloaded from http:/www.cambridge.org/core. University of Geneva, on 21 Nov 2016 at 12:36:55, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/S0075435816000617 174 PAUL SCHUBERT

that the general purpose was attained. In this particular case, the general purpose was to get every citizen of the Empire to declare a kind of allegiance to the traditional custom of sacrice, insisting both on past conduct and on present performance. The administrative infrastructure was in place and had served other purposes in the past; now Decius was able to use it for this Empire-wide sacrice.

II LIBELLI: AN OVERVIEW OF PAST RESEARCH

The Decian Persecution in the Christian Sources Until the end of the nineteenth century, our knowledge of the events mentioned above was limited to the testimony of three Christian writers: Cyprian, Bishop of Carthage, a contemporary witness who suffered personally from the consequences of the edict; Eusebius of Caesarea (c. A.D. 260–339) who, in his History of the Church, quoted the correspondence of Dionysius, Bishop of Alexandria at the time and a self-described victim of the imperial order; and the Passio Pionii, the account of a martyrdom that took place in Smyrna.12 Understandably, those Christians who suffered through Decius’ edict not only resented the treatment, but also claimed that this was a case of persecution directed specically at Christians. This took place in a context where the Empire was still resisting the fast growth of a religious phenomenon that would cause, a few years later, an undeniable persecution against the Christians (in A.D. 257 under Valerian and Gallienus). The measures taken by Diocletian in his own edict on sacrice imposed a much harsher treatment upon Christians, including conscation, prohibition, and arrest.13 In contrast, Constantine’s position towards Christians apparently evolved in several steps from mere tolerance towards full acceptance and conversion.14 The standard wording used by Christian authors has long been that of ‘Decius’ persecution’, and it has endured till this day in many instances.15

A New Perspective: Papyri A change of paradigm came about, however, with the discovery in Egypt of some certicates of pagan sacrice written on papyrus.16 The rst libellus was published in 1893: it is a Berlin papyrus dated 26 June A.D. 250, and originates from the village of Alexandrou Nesos in the Arsinoite nome.17 It was followed by another certicate from the Arsinoite nome in 1894, in the Vienna collection;18 in 1904, Grenfell and Hunt published the rst libellus from Oxyrhynchus.19 Those were the rst testimonia pertaining to the event that carried no explicit judgement on it. None of them provided any clear evidence of an action directed specically at Christians;20 but the idea of a persecution against a particular community could not be disproved either.

12 See Rives 1999: 136. 13 See Rives 1999: 142. 14 See Barnes 2007: 202. The degree to which Constantine can be considered a Christian emperor has been hotly debated; see Lenski 2006:9–10. 15 See e.g. Clarke 1969; Clarke 1984: 21; Luijendijk 2008: 155. The Latin word persecutio appears already in the opening paragraph of Cyprian’s De lapsis. The Greek word διωγμός is used in the Passio Pionii 2.1 and in Eusebius’ History of the Church 6.39.1. 16 See Knipng 1923: 345–6. For a good overview of the way in which the Christian texts t together with the newly discovered papyri, see Clarke 1984:25–30. 17 See 1 = BGU I 287 = W.Chr. 124. 18 See 9 = SB I 4455. 19 See 7 = P.Oxy. IV 658. 20 The names of the applicants were examined by Knipng 1923: 359–61, who could not nd any clues pointing to their being Christian; see also Keresztes 1975: 763 and Luijendijk 2008: 173. In A.D. 256, an order for arrest from Oxyrhynchus species that the wanted person is a Christian, which implicitly suggests that this is

In 1907, yet another certicate, stored at the (modern) Museum of Alexandria but originating from Ptolemais Euergetis in the Arsinoite nome, radically changed the perspective on two counts.21 First, a three-digit number at the top of the document (υλγ = 433, damaged but convincingly legible on the photographic plate) suggested that this papyrus had been part of a τόμος συγκολλήσιμος, i.e. a roll made of separate certicates glued together for the purpose of recording by the administration. This implied that, in the nome capital, at least 433 individuals or households had undergone the required performance of a sacrice. The presence of so many Christians in Ptolemais Euergetis in A.D. 250 seems hardly likely, especially as Christians did not leave other notable traces in the papyri from that area at such an early date. The second element was of even greater importance: the individual who was performing the sacrice happened to be an Aurelia Ammonous, daughter of Mystes, priestess of the Egyptian god Petesouchos. Here was an unmistakable element that went against the idea that Decius’ edict was directed precisely against the Christians. The whole idea of an intentional persecution against a denite religious community was put into question. Several historians concluded that the edict was aimed at every individual in the Empire, regardless of his religion. When another isolated certicate was published in the same year, 1907, one could not yet tell that this would be the tip of — as it were — a large papyrological iceberg:22 for it later became clear that this document, written in the village of Theadelphia, belonged to a larger group of some further nineteen libelli, purchased for the most part by the Library of Hamburg and published by Paul Meyer in 1910.23 Several other libelli from Theadelphia turned up in other collections in the following years, bringing the total of the documents from the so-called ‘Decian libelli archive’ from Theadelphia to thirty-four pieces — to which could be added four uncertain cases.24 In what follows, this archive will be important both because it offers a large group of homogeneous documents that allow internal comparison, and because it is also possible to compare these papyri with those found outside of Theadelphia.

uncommon; see P.Oxy. XLII 3035, 4–5: Πετοσορᾶπιν Ὥρου χρησι|ανόν (l. χριστιανόν); Luijendijk 2008: 174– 84. On the irregular spelling χρησιανόν, see Luijendijk 2008: 140, as well as J. Rea in P.Oxy. XXXVI 2785.2, commenting on SB XII 10772.10 (= PSI XIV 1412 = SB VI 9451; mid-/ late third century), where he proposes to read δι<ὰ> Σώτου τοῦ χρησ<τ>ια[νοῦ]. 21 See 3 =W.Chr. 125. This papyrus was known to a few scholars in 1900, but was formally published only in 1907; see Knipng 1923: 346, n. 6. 22 See 27 = SB I 4445. This is the document mentioned by Ulrich Wilcken as ‘einer aus Philagris im Faijûm’ in his introduction to W.Chr. 124. See SB I 4445.3–6: ἀπὸ | κ̣ώ̣μης Φιλαγρίδος κατα|[μέ]ν̣ουσα ἐν κώμῃ Θεα|[δελ]φεί̣ᾳ ‘(…) from the village of Philagris, residing in the village of Theadelphia’. Philagris was located by Grenfell and Hunt, in 1907, in the vicinity of Euhemeria (therefore close to Theadelphia); see P.Tebt. II, app. II, § 5, s.v. Φιλαγρίς. 23 See Meyer 1910. 24 For an overview, see , which provides a list of the relevant documents. They correspond to numbers 10 to 43 in the list provided at the end of this article. The case of 1 = BGU I 287 = W. Chr. 124 remains uncertain: it originates from Alexandrou Nesos, in the district of Themistos, where Theadelphia is also located; the document, however, does not t well into the pattern of the documents found in Theadelphia. The same applies to 2 = P.Ryl. I 12 (Ptolemais Euergetis); 4 = P.Lips. II 152 (the libellus published last, in 2002; Euhemeria); 5 = SB VI 9084 (Euhemeria). Regarding the last two cases: Euhemeria is fairly close to Theadelphia, but in those two documents the commissioners are not the ones who appear in the Theadelphia certicates. By contrast, one document included in the Theadelphia archive (34) also comes from Euhemeria, but the commisioners are from Theadelphia and the scribe has indicated explicitly that the person resides in Theadelphia. Another case is 33 = SB I 4439.3–6: π(αρὰ) Αὐρηλίου Ὡρίωνος | Κιαλ̣ῆἀπὸ κώμης | Ἀπιάδος καταμένων (l. -μένοντος)|ἐν κώμῃ Θεαδελφείᾳ ‘from Aurelius Horion, son of Kiales, from the village of Apias, residing in the village of Theadelphia’.

The Context of Discovery of the Theadelphia Papyri In the season 1898/9, Bernard Grenfell and Arthur Hunt excavated the site of Theadelphia (modern Batn el-Harît, also called Kôm Ihrit).25 Then in 1902, the Deutsches Archäologisches Institut also searched the same area.26 Grenfell and Hunt found a few papyri; so did the German excavation team. It seems that the local workmen continued to dig on their own after the European visitors had left. In all likelihood, the Theadelphia libelli archive was thus discovered in the form of a single batch in the early twentieth century. Many documents were complete; others were damaged, and preserved the bottom part of the papyrus sheet, often only with the date.27 This suggests a pile of papyrus leaves, with either the top or the bottom of the pile damaged at the time of nding. In one striking instance, two certicates display a pattern of holes that overlap very closely; they must have been squeezed together at the time when the holes were made.28 As far as can be ascertained from the available plates, the papyri were not folded (as opposed to for example 6, a document from Narmouthis, where the breaks on the sheet clearly indicate that it was folded vertically and horizontally). The bulk of this ensemble found its way, rst to the antiquities market, then to the collection of the Hamburg University Library; a few scattered pieces were purchased by other buyers.29 The certicates found outside of Theadelphia do not belong to an archive and are for the most part unrelated one to another. Apart from the libelli archive, the unofcial foraging for papyri on the site of Theadelphia yielded two other important nds. The rst is the Heroninos archive, belonging to an estate-manager (φροντιστής) in Theadelphia from September 249 till Summer 268.30 The papyri were found ‘partly during excavations in the period 1899– 1913 and partly by native diggers about 1900–1903’.31 We shall see below that two individuals from the Heroninos archive appear also in the libelli. The Sakaon archive came to light in the same context: part of the ensemble ‘entered the Egyptian Museum in Cairo in spring 1903’.32 The texts date from the third to fourth century and seem unrelated to the libelli archive.

Interpreting the New Setting Since Knipng’s seminal article in 1923, several major contributions have been made by other scholars.33 The discussion has focused, for the most part, on two issues: whether Decius’ edict was directed specically at Christians or whether it was aimed at all citizens of the Empire;34 and what overall intent lay behind this extraordinary measure.

25 See P.Fay. p. 54–62; Rathbone 2009: 18; Leuven Homepage of Papyrus Collections, archive no. 331. 26 See Rubensohn 1902:47–8; Primavesi 1996: 178, n. 31. 27 For libelli where the bottom part is preserved, see 5, 11, 13, 14, 18, 19, 20, 24, 28, 29, 31, 37. An exception is to be found in 41, where only the heading is preserved. 28 These are 10, a papyrus from the Hamburg collection, and 21, belonging to the Rylands collection. 29 Collections other than Hamburg: Berlin (16), Florence (32), Rylands/Manchester (21, 25, 26, 36, 42), Vienna (27). Two documents (17 and 22) were acquired as late as 1920 for the Michigan collection. 30 See Rathbone 1991 and 2009. 31 See Leuven Homepage of Papyrus Collections, archive no. 103. Grenfell and Hunt, as well as the German team, already found a few pieces from this archive. Most documents were acquired on the antiquities market, either on behalf of the newly founded Società Italiana per la Ricerca di Papiri, or by Carl Wessely, who bequeathed them to Theodor Hopfner; the latter took his lot to Prague, see Rathbone 2009: 21. 32 See Leuven Homepage of Papyrus Collections, archive no. 206. 33 For a good overview of research on the topic, see Molthagen 1975, especially 61–84, and Leadbetter 1982;ata more recent date, Luijendijk 2008: 157–9 and Brent 2010: 123–8. 34 The absence of any evidence regarding the Jews suggests that they were exempt; see Rives 1999: 138, n. 16.

Relying heavily on the testimony of Christian authors, Paul Keresztes believed that Decius was targeting specically the Christians.35 In his view, the libelli should be read in the light of literary sources (and not vice versa). Moreover, Robin Lane Fox was of the opinion that, if every single Roman citizen was supposed to perform a sacrice under the control of local authorities, this would have led to a bureaucratic nightmare;36 it would seem more plausible that only a small fraction of the population had to submit to this procedure. To this double-pronged approach, Reinhard Selinger and James Rives replied: by underlining the fact that the literary sources were written ‘by Christians for Christian audiences’ and had thus introduced a strong bias in the interpretation of the events; and by drawing a parallel with the Empire-wide census, a bureaucratic procedure for which the Romans had both experience and an adequate administrative infrastructure.37 Whereas Rives refrained from trying to explain the emperor’s true motivation, Allen Brent — drawing on Selinger — considered that Decius’ edict could be best explained as a kind of supplicatio in which every individual was required to seek the gods’ protection over the Empire.38 A consensus now seems to have emerged on the following points: (a) the administrative impact of Decius’ edict was relatively limited, both in terms of the involvement of ofcials (no more than in the regular census) and of duration (a few months in the spring and summer of A.D. 250); (b) although it does not seem that Decius was targeting the Christians in particular, he cannot have ignored that his order prescribing universal sacrice would have a deeply disturbing effect on the Christian community;39 (c) indeed, the Christians took this order as a measure directed specically against them, and this feeling was reected in their writings. More importantly, however, this edict was probably intended by the emperor as a way to unite the Empire in a search for the gods’ protection. Leaving aside the alleged restriction to Christians, we can also see that our certicates display no signicant distinction of gender (men and women are well represented in the extant libelli) or of age (children and old people appear too).40 This being said, we shall now turn towards the libelli found in Egypt, paying particular attention to the way in which the documents were drafted.

III SOME POINTS OF METHOD

It is a well-known fact that the 60,000 documentary papyri already published provide us with an unmatched corpus of sources for understanding the daily life of Graeco-Roman Egypt. In this huge body of evidence, forty-six libelli seem a small portion of the whole. They nevertheless allow for a most interesting case study for the following reasons: – The corpus is homogeneous, with documents dated in a period of only two months (Payni and Epeiph of Decius’ rst regnal year, corresponding roughly 41 to June–July A.D. 250).

35 See Keresztes 1975. 36 See Lane Fox 1986: 460–8. 37 See Selinger 1994:32–5 and 112–19; Rives 1999: 140 and 147–51. 38 See Brent 2010: 192: ‘Decius was thus celebrating an apotropaic supplicatio, whose aim was to produce the pax deorum at a particular stage in the metaphysical decline of the cosmos, particularly associated with the cultic ofce of the emperor himself and his dead and deied legitimate predecessors.’ To be fair, one should note that supplicatio is already mentioned by Rives 1999: 146. 39 See especially Corcoran 2015:69–71, with an excellent overview of the question. 40 In using the masculine to designate globally those — women and men — who had applied for a libellus, I only defer to a grammatical convention. Luijendijk 2008: 160, n. 24, underlines the fact that ‘a majority of libelli was issued to women (…)’. With a proportion of seventeen women against thirteen men, the statistical sample seems too small to recognize a signicant imbalance between genders. 41 Decius assumed the imperial title in the autumn of 249; he died in June 251. See Kienast 1996: 204.

– It includes a core of thirty-four papyri orginating from the same village, Theadelphia in the Arsinoite nome. Several scribes were at work, some of whom wrote more than one document. – It allows comparison with libelli found either elsewhere in the Arsinoite nome or in the Oxyrhynchite nome. – Numerous images are easily available, especially of the better preserved specimina. – The event that caused the writing of these documents is known through external sources which, however biased, offer another angle of view on the subject.

Both the literary and papyrological sources on the present topic have been closely examined by many scholars over more than a century. In the meantime, however, it has become relatively easy to constitute a stock of images of any given kind of documentary papyri.42 Some were published in print and are recorded in databases; others have now become accessible as digital images. For the present inquiry, we shall not need high-denition images because the focus will be set on the layout of the documents, and not on the precise reading of the texts. An examination of the images should, however, allow us to improve our knowledge of the procedure initiated by Decius. The approach followed here could easily be expanded to other similar corpora of papyri, with the general aim of better understanding how the documents were, in a very concrete way, written by the scribes. The wealth of available papyri makes it not only possible, but also necessary, to work in a selective fashion, since examining the original of each single papyrus in every library has become impossible due to the amount of available material. Some details might escape our attention, but this will not distort the overall picture. Incidentally, the case of the libelli also indicates that the images of papyri often display the best preserved specimina; the documents that are not easily accessible are for the most part very incomplete, often displaying only the date at the bottom of the certicate. In other words, a kind of natural selection took place; this process does not result from any co-ordinated effort on the part of papyrologists. Thus, instead of starting from the testimony of our Christian literary sources and moving towards the texts of the documentary papyri, I propose to focus on the available images of preserved libelli in order to understand the work of the scribes and their relation to the people who had to submit to Decius’ edict; only then shall we ask what was the precise purpose of the certicates issued by the local administration of Theadelphia and other villages.43

IV FORMAL ASPECTS

Grouping the Papyri The brief overview of research sketched above has already made clear one essential point: the corpus under consideration consists of a central core corresponding to a small archive

42 Available images are for the most part recorded in the Heidelberger Gesamtverzeichnis der griechischen Papyrusurkunden Ägyptens aswellasonpapyri.info . Some images that do not appear in these databases can nonetheless be found easily on the websites of libraries and museums. 43 My approach is diametrically opposed to that followed by Keresztes 1975: 775: ‘It would seem that only G. Schoenaich still maintains that the Decian edict was anti-Christian not only in spirit but also in word. In other words this quite general change of opinion, earlier based on Eusebius (Dionysius), Cyprian, and others, was the result of a reinterpretation of these same literary sources on the basis of the Egyptian libelli. Instead, perhaps, these Egyptian libelli should be interpreted in the light of the likewise contemporary reports of Dionysius and Cyprian.’

found in Theadelphia (thirty-four pieces), and of scattered documents that were found in other contexts (the remaining twelve libelli), either elsewhere in the Arsinoite or in the Oxyrhynchite nome. As we shall see, the layout of the libelli varies according to the place of writing; not surprisingly, the Oxyrhynchus documents display the most important variation in layout, whereas the libelli from the Arsinoite nome outside of Theadelphia bear a closer resemblance to the Theadelphia certicates.

The Format of Libelli in Theadelphia All the preserved libelli correspond to a well-dened type of document, namely a request submitted to an ofcial authority by an individual. For this purpose, the regular format was that of the ὑπόμνημα (lit. ‘memorandum’), which normally includes four fundamental parts: (a) heading; (b) description of the general purpose of the procedure; (c) specic request; (d) date. Starting with this very general outline, we should examine more closely the format of the libelli from Theadelphia because there is very little variation in the layout of these documents. This will make it easier to characterize the differences to be found in the libelli from other places in the Arsinoite or Oxyrhynchite nomes. The overall structure of the libellus in Theadelphia can be summarized as follows:

TABLE 1 Section 1 (1st hand) a) Heading: the document is addressed to the ofcials chosen to supervise the sacrice (τοῖς ἐπὶ τῶν θυσιῶν ᾑρημένοις), followed by the name of the applicant. b) General purpose of the procedure: the applicant states that he has always sacriced and that, in the presence of the ofcials, he has again sacriced and tasted the offering. c) Specic request: the applicant asks the ofcials to sign the document. Section 2 (2nd hand) The ofcials — two of them in the case of Theadelphia — state that they have seen the applicant sacricing. Section 3 (3rd hand) Actual signature of one of the two ofcials. Section 4 (1st hand) d) Date.

Several elements point towards a general pattern. In Theadelphia, all four fundamental elements of the ὑπόμνημα (parts a/b/c/d as listed above, corresponding to Sections 1 and 4 of the libelli) are written by scribes who — judging from the uency of the writing — are always professionals. Paul Meyer wrote of a ‘berufsmässiger Urkundenschreiber (νομογράφος)’.44 Such a technical label is not warranted by the documents and may seem slightly misleading. We can nonetheless recognize here the same kind of scribes who prepared census declarations, as described by Bagnall and Frier: ‘The very high degree of formulaic character found among declarations from a particular area and period would have been impossible without the dissemination of ofcially sanctioned standard forms, and it is unlikely that anyone except a professional would have had any means of learning these standard document types. Moreover, the handwriting of the bodies of the declarations is consistently professional, generally a fast cursive. (…) The contrast with the signatures (…) is often striking.’45

44 See Meyer 1910: 25. For an example of a Theadelphia libellus, see Fig. 2 at the end of this article. 45 See Bagnall and Frier 1994: 18.

There are thus different scribes at work, but their task is invariably the same: writing Sections 1 and 4. To this basic format, however, are added two specic elements between (c) and (d), written by another hand (Sections 2 and 3). Obviously, the certicates were prepared by a trained scribe before the applicant came into the presence of the commission in charge of supervising the sacrice, where the certicate was completed. The documents left a blank window between the request and the date; the ofcials could then ll in the window with the statement conrming proper sacrice and the commissioners’ signature. This element, seemingly trivial, will be of some importance when we compare the format of the Theadelphia libelli with some requests of another kind. In every case where an image is available, the statement conrming proper sacrice (Section 2) is written by the same hand.46 This scribe is relatively competent, but his training compares unfavourably with the writing of most of those who prepared Sections 1 and 4 of the documents. His writing is small, cursive, and it is conned between two virtual horizontal lines: letters such as iota or rho do not stick out markedly. On the whole, however, the ductus is rather irregular. One cannot exclude that this task was performed by Aurelius Serenus, one of the two ofcials in the commission, but it seems more likely that the ofcials received professional assistance for this task. Whoever repeatedly wrote this sentence did not systematically consider the gender of the applicant: in several cases where the heading indicates a female applicant, the scribe used a standard wording with the masculine participle θυσιάσοντα (l. -ζον-).47 The ofcial who signed the document (Section 3), Hermas, is probably the village scribe (κωμάρχης) known also from the Heroninos archive.48 In some cases, the signature is missing; and we do not nd any other signature in the extant Theadelphia libelli. Hermas’ clumsy hand is easily recognizable in several documents: each letter, in a large size, is formed separately and painstakingly; the ductus is irregular. Hermas belongs to the category of the so-called βραδέως γράφοντες, i.e. slow writers with a very limited command of the craft.49 He also abbreviates his signature, either to spare himself some effort or because his letters are so large that he lacks the space for writing his signature in full: ΕΡΜ ̅ΣΕΣΗΜ̅= Ἑρμ(ᾶς) σεσημ(είωμαι) ‘I, Hermas, have applied my signature’. Coming back to the various professional scribes who wrote Sections 1 and 4 of the Theadelphia libelli, it is possible to group the documents according to the hands, at least in the case of papyri for which an image is easily available. This classication was already made by Meyer, who recognized that each scribe, although agreeing on the general wording of the certicate, introduced some minor differences in the model he used.50 Each formulary relies on a given model and can be associated consistently with a particular hand. Meyer was able to observe this on the original papyri in Hamburg, including in cases where images are not easily available today, and the publication of more certicates after Meyer’s time has conrmed the close correspondance between the scribes and their specic models.

46 Here I disagree with Meyer 1910:27–8, who identied several hands at work. 47 See e.g. 15 (= SB I 4440.16–17): Αὐρήλιοι Σερῆνος καὶἙρμᾶςεἴδαμέ(ν) σε θυσιάσοντα (read θυσιάζουσαν) ‘We, the Aurelii Serenus and Hermas, have seen you sacricing’. 48 See Rathbone 1991: 20, n. 25. SB XX 14645.38 (= VI 9409; Theadelphia, A.D. 251): δι’ Ἑρμᾶ κωμάρχ(ου). Another document, P.Flor. III 345 (mid-third century A.D.), is a letter addressed to Hermas in his function as village scribe. 49 See Youtie 1971. 50 See Meyer 1910:25–7, where he distinguishes six different models/hands (A–F). In 1916, he identied a seventh (G) in P.Meyer 15 (see his introduction to this papyrus, p. 76).

TABLE 2

1 NUMBER REFERENCE MEYER MODEL/HAND MEYER DATE HEIGHT WIDTH 12 SB I 4436 2 A Payni 20 21 8.5 13 SB I 4437 3 A Payni 20 15 SB I 44402 6 A Payni 22 21 6 16 SB I 5943 6a A Payni 22 20.6 6.6 19 SB I 4442 8 A Payni 25 20 SB I 4443 9 A Payni 25 28 SB I 4446 12 A Payni 29 22 7.5 32 PSI V 453 A Payni 2- 20 7.5 42 SB III 6828 A ? no date 25 SB III 6827 B Payni 21 26 P.Ryl. II 112c B Payni 28 21.5 7.2 30 SB I 4448 14 B Payni 29 22 8 14 SB I 4438 4 C Payni 21 17 P.Mich. I 157 C Payni 23 20 5.7 18 SB I 4441 7 C Payni 23 22 P.Mich. III 158 C Payni 27 21.5 8.1 23 SB I 4444 10 C Payni 27 21.5 6 27 SB I 4445 11 C Payni 27 21.2 6.5 29 SB I 4447 13 C Payni 27 33 SB I 4439 5 C Payni - 21 6.5 36 P.Ryl. II 112b C no date 38 P.Meyer 17 20b C no date 39 SB I 4451 17 C no date 40 P.Meyer 16 20a C no date 11 P.Hamb. I 61a3 D ? Payni 19 24 P.Hamb. I 61b D Payni 27 31 SB I 4449 15 D Payni 29 35 SB I 4450 16 D Epeiph 20 25 6 43 SB I 44524 18 D no date 10 SB I 4435 1 E Payni 18 21.5 6 41 SB I 4453 19 F no date 34 P.Meyer 15 15a G Epeiph 3 20 9 Continued

Table 2 Continued 1 NUMBER REFERENCE MEYER MODEL/HAND MEYER DATE HEIGHT WIDTH 21 P.Ryl. II 112a Payni 26 18.5 11.9 37 SB I 44545 20 no date

1 Height and width are provided only for complete documents. 2 15 and 16 are duplicates. 3 Meyer, in the publication of P.Hamb. I 61a, suggests that this fragment belongs perhaps to the same document as 35. 4 43 was written for a woman from Theoxenis; the scribe did not specify that she lived in Theadelphia. The certicate, however, follows model D so closely that it belongs without any doubt to the group and was produced in Theadelphia. 5 This item is too small a fragment to allow a classication.

Model/hand A includes a pair of duplicate documents: not only were 15 and 16 written by the same hand, they were also produced on behalf of the same person, Aurelia Charis. This, added to the fact that some certicates do not display Hermas’ signature, indicates that the Theadelphia archive does not contain libelli which were in the hands of the applicants, but one or more copies that had been kept together by the village scribe. This group is also of interest because, although the certicates were prepared by the same hand, the scribe put uneven care into preparing the documents. In 12, 15 and 16, the cursive writing is rather elegant and regular. In 32, it is the same hand, but one notices that it runs much faster and carelessly. Arthur Hunt also recognized the fast hand of 21 as the same as that of 15;51 this last identication, however, will be reassessed below. Applicants needed professional scribal help for preparing their request. There were several scribes in the village who mass-produced the certicates in advance. The consistency in the format of the papyrus sheets makes this even more likely, as was already noted by Jacques Schwartz: ‘(…) la hauteur uniforme de presque toutes les déclarations de Théadelphie donne à penser que les diverses “premières mains” travaillaient pour un entrepreneur qui leur fournissait ces rectangles de papyrus, de proportions inhabituelles, après les avoir taillés dans des volumina;lanesse de fabrication de certains papyrus contenant ces certicats vient à l’appui de cette hypothèse.’52 In this respect, the consistency of the format is as striking as the few exceptions to be found in the batch. Let us focus on complete documents, leaving aside those that do not allow for precise measurement of height and width. Schwartz was right to underline the almost constant height of libelli, which ranges between 20 and 22 cm (see Table 2 above); the width is more irregular, ranging normally between 5.7 and 8.5 cm. Rolls were clearly produced at a standard height, from which narrow strips could be cut to produce the certicates. There are, however, three documents that do not t the standard pattern. One is exceptionally wide and short: 21 (11.9 by 18.5 cm); another is unusually wide: 34 (9 cm); and the last is taller than the rest: 35 (25 cm). As mentioned above, 21 was attributed by Hunt to the same hand that also wrote 15. The overall style is indeed the same, and some letters (e.g. beta of εὐσεβοῦσα) display a striking similarity between the two documents. One signicant difference lies in the way

51 See P.Ryl. II, p. 94. 52 See Schwartz 1947: 367.

the scribe wrote the pi of παρά in the heading. Whereas 15 (as well as 12, 16 and 32) displays an arched pi, the same letter in 21 was drawn in three strokes. When one turns to the wording of the formulary used, there are also some small differences that indicate that Model A was not used to write 21.53 Therefore, the unusual width of the leaf used to produce 21 can be explained by the fact that this is not the same scribe who wrote 12, 15, 16 and 32: he may be close to the scribe A, but he has introduced his own peculiarities into the process, and he cut his leaf to dimensions that differ markedly from the rest of our libelli from Theadelphia. 34 was produced by a very competent scribe. His writing is elegant and regular, with letters well separated; it displays some variation in the thickness of the strokes (shading); some letters are decorated by end-loops (serifs; e.g. kappa in line 7); the scribe makes a more frequent use of diairesis than his colleagues. He has also added a clause that is not found in other libelli from Theadelphia: instead of the standard ἀξιῶὑμᾶς ὑποσημιώσασθαι ‘I ask that you sign’, he wrote the more elegant διὸἐπιδίδωμι̣̣| ἀξιοῦσα ὑμῶνλαβεῖντὴν̣| ὑποσημίωσιν (12–14; l. -σημείω-) ‘Therefore I submit (the request) and ask you to receive the signature’. The certicate was not signed by Hermas: there is only the regular statement prepared by his assistant. Between the end of the request and the assistant’s statement, another hand has added, in a rather cursive writing, that the applicant resides in Theadelphia (although she originates from Euhemeria).54 To sum up, the unusual width of this libellus is to be explained presumably by the fact that we are dealing here with a particular scribe who was trained in the preparation of literary texts. Judging from the quality of the writing and the particular wording, he may well normally have been employed in the production of books, and not of administrative documents.55 Or else we can draw a parallel with a letter from the Heroninos archive that was written in a script normally found on literary papyri and that includes a Homeric quotation.56 Dominic Rathbone (himself quoted by Jean-Luc Fournet) notes: ‘Possibly the central secretariat [sc. of the administration of Appianus’ estate] included one or two scribes who were partly employed as literary copyists (…).’57 35 (Model/hand D) also seems to be a special case, but for different reasons. With a date of Epeiph 20, it is much later than the other certicates found in Theadelphia: all were produced in the month of Payni, with the exception of 34 (Epeiph 3). It seems that 35 was issued almost as an afterthought towards the end of Epeiph. Here the scribe was careless in his writing, presumably because he had most of the work behind him; he had already drafted 24, 31, 43 and perhaps 11. This carelessness is also mirrored in the writing of the scribe who made the statement of proper sacrice: his hand is easily recognizable, but is nonetheless more irregular than in the other certicates where his statement appears. Besides the sloppy writing, 35 contains another interesting feature: it was issued on behalf of Aurelius Euprodokios, a labourer in the service of Aurelius Ap(p)ianus, the owner of the large estate covered by the Heroninos archive.58 Since the libellus landed

53 A διατετέλεκα / 21 διετέλεσα;Aἔσπισα καὶἔθυσα / 21 ἔθυσα καὶἔσπισα /Aὑποσημιώσασθαί μοι / 21 ὑποσημιώσασθαι. 54 See 34 = P.Meyer 15.17: καταμένων (l. -μένουσα) ἐπὶ κό(μης) (l. κώ-) Θεαδελφίας. 55 There are fourteen literary papyri listed in the Leuven Database of Ancient Books with a provenance from Theadelphia and a date in the second/third or third century. Among these, I could nd none that was copied by the same hand as 34. 56 See Fournet 2012: 141–2. To the examples of the use of Homeric material in a documentary context listed by Fournet can be added the case of P.Nicole inv. 72, a fragment of a commentary on the Iliad which was recycled into a business note for an estate, presumably of the same type as that of Appianus; see Schubert 2007: 139–42. 57 See Rathbone 1991: 12. 58 Euprodokios is an οἰκέτης, a function that applies to an individual in a position of dependance, without necessarily being a slave; see Rathbone 1991: 106–16.

in the Hamburg collection, it presumably belongs to the Theadelphia libelli archive, not to the Heroninos archive. There is in fact no explicit mention of Euprodokios in the Heroninos archive although the wording clearly relates him to the Appianus estate.59 To conclude on these various points, applicants needed professional scribal help to prepare the request. There were several scribes in the village who mass-produced the certicates in advance; the consistency in the format of the papyrus sheets makes this even more likely. The closeness of the various models (A to G) suggests that the same model was provided to the scribes, but that each one made some small adjustments when he took over the model. Although most of the scribes followed a standard procedure, the format of the papyri mirrors a special treatment in some particular cases (21, 34 and 35). Once the applicant had his request prepared, it was submitted to the commission, presumably in more than one copy. The Theadelphia archive corresponds to a batch of libelli that remained in the hands of a village ofcial. The applicants must have kept a copy for themselves.

Other Libelli from the Arsinoite Nome Several libelli were found that also originate from the Arsinoite nome, but from towns or villages other than Theadelphia. They do not belong to the archive that was discussed above.

TABLE 3

1 NUMBER REFERENCE MEYER DATE PROVENANCE HEIGHT WIDTH 1 BGU I 287 21 Epeiph 2 Alexandrou Nesos 20.5 8 2 P.Ryl. I 12 25 Payni 20 Ptolemais Euergetis 22.0 12.4 3 W.Chr. 125 23 no date Ptolemais Euergetis 4 P.Lips. II 152 Payni 22 Euhemeria 20.5 7 5 SB VI 9084 Payni 23 Euhemeria 6 P.Wisc. II 87 Payni 10 Narmouthis 21.0 7.8 9 SB I 4455 22 no date Philadelphia 45 PSI VI 778 Epeiph 2 Arsinoite nome (?) 1 Height and width are provided only for complete documents.

The scribes seem to rely on a model that is basically the same as in Theadelphia, although the scribal habits can vary. The size of the leaf is identical with the Theadelphia standard (see table above), with a height between 20.5 and 22 cm, and a width that varies from 7 to 8 cm.60

59 See 35 = SB I 4450.3–7: παρὰ Α̣ὐρ̣ηλ(ίου) Εὐπροδοκί̣ου οἰκ<̣έ>του Αὐρηλ(ίου) Ἀπιανοῦἐξη(γητεύσαντος) τῆςλα(μπροτάτης) πόλ(εως) τῶν Ἀλε̣ξ(ανδρέων) κ(α)ὶὡς̣χρ̣η(ματίζει) κα̣τα̣μ̣έ̣ν̣ων̣(l. -μένοντος) ἐν ̣̣ Θεαδελφίᾳ ‘from Aurelius Euprodokios, labourer of Aurelius Appianus, former exegetes of the most illustrious City of the Alexandrians (and other titles), residing in (…) Theadelphia’. 60 For an example, see Fig. 1 at the end of this article.

The only certicate that was produced in a different size is 2 (height 22 cm / width 12.4 cm), which again may be a special case. It was written in Ptolemais Euergetis, the capital of the Arsinoite nome, by Aurelius Eirenaios, on behalf of his wife Aurelia Demos, who does not know how to write. Eirenaios himself is a skilled writer; he uses an elegant cursive script. His writing is to be found also in P.Flor. I 19 (Ptolemais Euergetis, A.D. 248), where he offers to take on lease some land owned by a Council member of the nome capital.61 It thus seems that Aurelius Eirenaios, who is not necessarily a professional scribe, nonetheless had sufcient skills to produce his own certicate, presumably following an available model. He cut his papyrus sheet in a size that is close to that of P.Flor. I 19, but differs from the standard width to be found among professional scribes who mass-produced the libelli.

TABLE 4

2=P.RYL. I 12 P.FLOR.I19 Height 22 21 Width 12.4 11

The libelli are structured in the same way as in Theadelphia, with Sections 1 (parts a/b/c of a ὑπόμνημα) and 4 (part d = date) written by the scribe who prepared the document in the rst place. A space is also left between Sections 1 and 4 for certication by an ofcial. In three instances (1, 2, 9), the applicant states explicitly that he has submitted his request, using the phrasing ἐπιδέδωκα ‘Ihavesubmitted’, which is found commonly in ὑπομνήματα but never appears in the libelli from Theadelphia. In 2 and 9, the applicants are illiterate and the submission of the document is made by someone else in their stead. Outside of Theadelphia, scribes from the Arsinoite also tend to include some information that allows the identication of the applicants (age, scars; see 1, 2, 4, 6). The notice of certication written by the commissioners’ assistant (Part 3 in the structure of the Theadelphia libelli) regularly appears, but the actual signature of a commissioner (Part 2) is scarce: only in one instance (6), six commissioners sign individually. To sum up, the scribes from various places in the Arsinoite nome seem to rely on a basic model similar to the one in use in Theadelphia. The model can display some small differences, of the same scale as those found in Theadelphia, where each scribe introduced his own little habits (spelling mistakes, preference for a tense, etc.). The main differences do not relate to the model that was provided to the scribes, but have to do with scribal habits in a given village. The overall impression remains one of great consistency, not only in Theadelphia but over the whole Arsinoite nome. One can assume that scribes in all villages of the Arsinoite nome were issued the same basic model.

Libelli from the Oxyrhynchite Nome: a Comparison with the Certicates from the Arsinoite Apart from the thirty-four libelli that constitute the Theadelphia archive, and from the other eight that were found in other villages in the Arsinoite nome, there are four certicates from the Oxyrhynchite nome.

61 Image: P.Flor. I, Tav. VI (where the papyrus was mislabelled as P.Flor. I 21; correct to 19), to be compared with 2 = P.Ryl. I 12, pl. 12.

TABLE 5

1 NUMBER REFERENCE MEYER DATE PROVENANCE HEIGHT WIDTH 7 P.Oxy. IV 658 24 Payni 20 Oxyrhynchus 8 P.Oxy. XII 1464 Epeiph 3 Oxyrhynchus 17.2 9.8 44 P.Oxy. LVIII 3929 Epeiph – Thosbis 46 P.Oxy. XLI 2990 no date Oxyrhynchus 1 Height and width are provided only for one complete document. Although the three remaining papyri are incomplete, the width can be estimated fairly securely at 7 cm (7) and 7.5 cm (44 and 46).

The density of nds in the Arsinoite and Oxyrhynchite nomes does not imply that libelli were not produced in other nomes of Middle Egypt, or in the Delta or in Upper Egypt. The Arsinoite and Oxyrhynchite nomes have yielded the largest quantity of papyri overall, and if it were not for the Theadelphia archive, our corpus of libelli would be much more limited. The rst three Oxyrhynchite certicates (7, 8, 44) are preserved almost entirely. Regrettably, the papyrus breaks off at the bottom of 7 and 44, whereas 8 seems complete but lacks the requested certication and signature of a commissioner. We encounter the reverse situation in the fourth libellus (46), which preserves only a statement — without proper signature — of two commissioners, each in his own hand; the writing of the commissioners is fast and cursive, but does not necessarily indicate the work of a professional scribe.62 A comparison between the documents from the Oxyrhynchite nome and those from the Arsinoite brings to light both similarities and differences. First, the Oxyrhynchite wording echoes the Arsinoite in many places, e.g. ἀεὶ μὲνθύωνκαὶ σπένδων [τοῖ]ςθεοῖς̣̣[δ]ιετέλ[εσα (7)/ἀεὶ μὲνθύεινκαὶ [σπέ]νδειν καὶ σέβ̣ειν̣ θεοῖςεἰθισμένος (8)/ἀεὶ μὲνθύωνκαὶ σπένδων τοῖςθεοῖςδιετέλουν(44) ‘I have constantly sacriced and poured libations for the gods’. See also ἀξιῶ (ὑμᾶς) ὑποσημιώσασθαι μοι (7, 8, 44) ‘I request that you sign for me’. Whereas certicates from the Arsinoite nome follow several models that display only minor differences among themselves, the Oxyrhynchite libelli look less homogeneous in their phrasing. The address is notably inconsistent: τοῖς ἐπὶ τῶν ἱερῶν [καὶ] θυσιῶνπόλ[εως] (7)/[τοῖς] ἐπὶ τῶνθυσιῶναἱρεθεῖσι τῆς [Ὀ]ξυρυγχειτ̣ῶν πόλεως (8)/τοῖς ἀναδοθεῖσι̣̣ἐπὶ θυσιῶ(ν) κώμης Θώσβεως (44). The Oxyrhynchite certicates can also be consistent in their differences. For example, the Oxyrhynchite formula ἐνώπιον ὑμῶν (7, 8, 44) ‘in your presence’ mirrors precisely and regularly the Arsinoite wording ἐπὶ παρόντων ὑμῶν / ἐπὶ παροῦσι ὑμῖν. This difference cannot be explained by a scribal preference of one or another formula in either nome: neither ἐνώπιον ὑμῶν nor ἐπὶ παρόντων ὑμῶν appear in other administrative documents; they seem quite specic to the wording of libelli. Another striking difference between the Arsinoite and Oxyrhynchite certicates lies in the presence of a blank window in the Arsinoite certicates, where the commissioners applied their signature. None of the Oxyrhynchite libelli displays such a window. It is likely that their signature was written at the bottom of the sheet; but the preserved documents do not allow a rm conclusion on this. The only complete Oxyrhynchite libellus is both shorter and wider than the Arsinoite standard. The Oxyrhynchite certicates consistently display one difference with the Arsinoite documents which seems at rst glance trivial, but will require further discussion.

62 For an example, see Fig. 3 at the end of this article.

TABLE 6

ARSINOITE NOME OXYRHYNCHITE NOME past The applicant has always taken part The applicant has always taken part in conduct in sacrices (and shown piety sacrices (and shown piety (8)) and towards the gods (models A/B)). poured libations for the gods. present He takes part in a sacrice, pours He takes part in a sacrice, pours action libations, tastes the offering and libations, tastes the offering and requests that his action be certied. requests that his action be certied.

In the Arsinoite nome, libations are not mentioned in the past conduct, but only in the present action. In the Oxyrhynchite, libations belong to both past conduct and present action; in other words, the only difference between past and present conduct lies in the tasting of the offering. We shall come back to this below. To sum up, the overall picture that emerges from this comparison looks as follows. Starting in Rome, the imperial edict must have been transmitted to the provinces of the Empire. In Egypt, the Prefect sent a copy — in Greek — of the original edict to the strategoi in charge of the nomes, instructing them to apply the measure. The wording was quite specic because the content of libelli in the Arsinoite and Oxyrhynchite nomes looks very consistent, even if the wording differs in some specic clauses. Although the wording of the edict is lost, we can be fairly sure that it included the following elements: every citizen must present himself before a local commission; he must acknowledge a past conduct that conforms with the traditional practice of sacrice; he must perform the same actions before the commission. Another element that was most probably included was ingestion of the sacrice meat, to which we shall return. The precise layout of models used for the certicates was determined at the level of the nomes, according to the scribal practice. In the Arsinoite nome, the consistency of the format, not only in Theadelphia but also in other villages in two districts of the nome (Themistos and Herakleides), indicates that a model was issued for the whole nome, with a precise layout. In the Oxyrhynchite nome, as far as we can judge from the scanty evidence available, models were not as consistent as in the Arsinoite.

V EFFECTIVENESS OF THE CONCRETE PROCEDURE

Tasting the Offering The brief summary of past reseach has shown that, at least in the eyes of some scholars, the procedure involved here would have been a bureaucratic nightmare if it had been applied to all citizens of the Roman Empire and had not been specically directed at one group — Christians. This was answered by drawing a parallel with the census that befell all citizens at regular intervals, a procedure for which the administrative infrastructure already existed on a permanent basis. We shall come back to it later; at this point, however, it is necessary to ask what precise action the issue of certicates throughout the Empire implied. In other words: what was actually required of the applicants, and why did they have to submit an application, although they had asked for nothing in the rst place? The libelli describe the applicant’s past conduct, followed by specic action taken on the day when the certicate was issued. In the past: reverence towards the gods, sacrices, libations (this last element only in the Oxyrhynchite nome); presently: sacrice, libation and tasting of the offering. The list of actions performed before the commissioners does not, however, imply the same degree of participation for each action.

‘Taking part in the sacrice’ (ἔθυσα) does not mean that every person, including women, children and elderly people, held a knife and slaughtered a sheep; it is more likely that mere attendance was recorded here. Also, slipping away from such a ceremony must have been relatively easy.63 Pouring a libation is already more precise because it implies that a person holds a receptacle and lets out some liquid; but it allows a degree of bodily distance that is quite different from the third type of action, namely tasting the offering. This last part of the procedure requires that the person swallow a piece of meat, which will then be digested, becoming almost a part of the person. There is no way in which the applicant can keep any distance from this last element in the process. Dening participation through ingestion goes back to very ancient times: already in Greek myth, Persephone is attached for ever to the Underworld after she has swallowed a seed from a pomegranate.64 This is not to say that every villager in Roman Egypt would have remembered the Homeric Hymn to Demeter in detail; rather, the idea of swallowing food so as to indicate membership of a community is deeply rooted in an old tradition. Therefore, the symbolic value of tasting the offering could not pass unnoticed by most of the people involved. The formulary similarity between the declarations made in A.D. 250 and the alleged password of Eleusinian initiates should also be underlined. The standard phrasing in a certicate of sacrice is: ἔσπεισα καὶἔθυσα καὶ τῶν ἱερείων ἐγευσάμην καὶἀξιῶὑμᾶς ὑποσημειώσασθαί μοι ‘I have made a libation, taken part in the sacrice and tasted the offering, and I request that you sign for me’. According to Clemens of Alexandria, the password for the Eleusinian initiates was: ἐνήστευσα, ἔπιον τὸνκυκεῶνα, ἔλαβον ἐκκίστης, ἐργασάμενος ἀπεθέμην εἰς κάλαθον καὶἐκ καλάθου εἰςκίστην‘I have fasted, I drank the kykeon, I took from a chest, I worked and deposited into a basket, and from the the basket into the chest’.65 If Decius’ intent was indeed to offer a kind of supplicatio to the gods, the declaration made in the certicates could well offer a remote echo to the declaration of the performance of a religious ritual that had been taking place for many centuries. A parallel with modern practice, although imperfect, can also help to illustrate the point. A non-believer can attend Mass in a Christian church, for example on the occasion of a wedding or a funeral. Attending Mass does not make him a Christian; he does not even have to take part in prayer. On the other hand, when the priest distributes the Holy Communion, a non-believer will not join the line to swallow the eucharistic host — the equivalent of the sacriced animal’s esh — mixed with the wine that symbolizes the blood of the victim.66 The actual dividing line between those who belong to the congregation and others will be drawn, in most cases, at the moment when parish members line up to receive the Holy Communion. Likewise, in the case of pagan sacrices performed in Roman Egypt, the crucial point that must have really mattered in terms of the procedure was the fact that every person was obliged to taste the offering. In the certicates issued in the Oxyrhynchite nome, this point stands out very clearly: the only difference between the declaration of past behaviour and the present action lies in the tasting of the offering. This does not mean that, for Christians, a statement of past behaviour or the mere presence — i.e. passive

63 See Rives 1999: 147, commenting on Roman traditional religious practices: ‘(…) those who were simply uninterested could no doubt quietly absent themselves without anyone caring, and some people presumably did just that: such events were not to everyone’s taste’ [pun unintentional]. 64 See Homeric Hymn to Demeter 372; 393–7; 411–12. Richardson 1974: 276: ‘In general eating and drinking ratify one’s membership of a community (…).’ 65 See Clem. Alex., Protrept. 2.21.2; Foley 1999: 68. 66 See Matth. 26.26–8. The Holy Communion is acknowledged by the Apostle Paul as a way of partaking in the body and blood of Christ; see 1 Cor. 10.16. See also 1 Cor. 8.7: τινὲςδὲ τῇ συνηθείᾳἕως ἄρτι τοῦ εἰδώλου ὡς εἰδωλόθυτον ἐσθίουσιν, καὶἡσυνείδησις αὐτῶν ἀσθενὴςοὖσα μολύνεται ‘Some people, because they have been until recently acquainted with idols, eat (of the meat from the sacrice) as if it were offered to the idols, and their consciousness, being weak, becomes sotfened’.

attendance — on the occasion of a sacrice did not matter;67 but in comparison with simple participation, swallowing the esh from the sacrice was another matter because of the much stronger symbolic value placed in this ingestion. Christian writers similarly lay emphasis on the pollution incurred by the food they were forced to take.68 The Christian communion was, from its inception, a symbolic transposition of the pagan sacrice, where the slaughtered animal and its blood were replaced by vegetal elements, namely bread and wine.69 There is no way in which the authorities of the Empire could ignore this. Therefore, although the edict was not directed expressly against the Christians, the implicit consequence was that it would force Christians — among others — to perform a traditional sacrice instead of its symbolic transposition. The procedure that was decided upon, which included very specically the ingestion of the sacrice offering, forbade any kind of ambiguity: no one could later claim that he had merely attended a ceremony, but had not taken part in it.

Requesting a Signature A second element in the procedure deserves closer examination: although we can presume that nobody had asked to appear before a village commission in the rst place, it was nonetheless the duty of every individual to submit an application, asking that his participation in the sacrice be recorded and certied by the commissioners’ signature.70 The importance of this signature will become more evident when we compare libelli with census returns. Since the obligation of performing a sacrice presumably befell every citizen in the Empire, some doubts were raised about the administration’s capacity to organize such a procedure at relatively short notice. The parallel of the census returns shows, however, that an effective administrative apparatus was in place, at every level of each province, and that it could easily have been used also for the extraordinary procedure that was 71 initiated in A.D. 250. In fact, the Romans were able to manage comparable measures at short notice already in the rst century B.C., as is shown by the Heraclea Table: a general census could be ordered within a period of sixty days.72 To come back to the reign of Decius, a consistent chain of transmission was in place, from the highest level of each province to the smallest village, with ofcials who could forward orders from top to bottom and receive feedback from subordinate ofcials. Judging from the perspective of the province of Egypt, the parallel with the census returns shows that those ofcials were reasonably competent, even at the level of a village: in general, they knew how to use a model that was provided to them; they could delegate the reproduction of the model to independent scribes; and they had the capacity to ensure that virtually everyone in the village complied with the orders. In Egypt, there was — to

67 This point appears clearly in the pseudo-Clementine novel: see Clem., Rec. 2.71.3: ‘hoc enim pro certo scire te uolo, quia omnis qui idola coluit aliquando et eos quos pagani nominant deos, adorauit uel de immolatis eorum degustauit, spiritu inmundo non caret’‘For I want you to know this with certainty, because whoever has once worshipped idols and adored those whom pagans call gods, or has tasted of the offerings made to them, is not deprived of an impure spirit’. Although the narrative takes place in the reign of Nero, this novel was shaped between the second and fourth centuries A.D.; see Geoltrain 2005: 1186. Therefore the perspective adopted in this passage of the Clementina seems compatible with that of Christians in the reign of Decius. 68 See Rives 1999: 137, n. 13, who mentions the papyri in passing, focusing instead on the testimony of our Christian sources: Cypr., Ep. 31.7.1; Laps. 2 and 28; Passio Pionii 3.1. 69 See Popkes 1976, especially section VIII d: ‘Opfermahl’. 70 See Leadbetter 1982: 181. 71 On the chronological distribution of census returns, see Bagnall and Frier 1994:2–11. The bulk of our census returns date from the second century A.D. They rst appear towards the end of the reign of Augustus, and disappear after the census of A.D. 257/258, a few years after the reign of Decius. Some census records were being maintained as late as A.D. 267, and population lists endured into the 270s. Whether there is a correlation between Decius’ edict of A.D.250 and the disappearance of the census in the following years remains an open question. 72 See Legras 1907:31–2.

quote Clarke —‘by Roman standards, an exceptionally well-developed bureaucracy’, although ‘there was room for people to slip through the ofcial nets’.73 It should be noted, however, that the administration spent a lot of energy in bringing people back to the villages, especially those who had ed presumably to Alexandria because of the tax burden in the villages.74 There were no doubt some holes in the net, but they were small and Egyptian ofcials did their best to mend them continuously. In the process of producing the certicates, a considerable part of the work was actually outsourced to scribes who copied the forms for the applicants. The burden of preparing the documents thus lay principally upon the shoulders of the village’s inhabitants, not of the village scribe. It was the individual who submitted a request to the authorities, and not the reverse. Therefore the layout of the document, in the case of both the census return and the certicate of sacrice, was basically the same. It corresponded to a memorandum (ὑπόμνημα), the structure of which was described above (typical structure of a libellus in Theadelphia). We have seen, however, that the certicates of sacrice included some elements that do not appear in every kind of ὑπόμνημα. Those who produced a model for a particular procedure followed the basic structure of the ὑπόμνημα and added some specic elements corresponding to the action required. At this point, a comparison with census returns is only partly adequate: for the census indicates that the administrative apparatus was available, but the models used for census differ from the libelli, both in form and purpose. In this comparison, therefore, we should add a third kind of document, declarations of death.75 Declarations of death can be associated with census returns because they share a similar function, from a distinct perspective: census returns allow village scribes to establish a register of all taxable inhabitants; death certicates are required in order to remove a person from the register, so that his family will not be liable to the poll tax in his stead. This relation was already underlined in passing by Rives who, however, did not elaborate on the form of the documents.76 The census returns differ from the other two types of documents notably in that they follow a one-way process: the applicant submits a declaration (ἀπογραφή) but does not request anything in return. E.g. SB XVIII 13289.14 (Theadelphia, A.D. 161): διὸ ἐπιδίδωμι τὴν ἀπογραφήν ‘Therefore I submit the declaration’. Once he has made his declaration, his obligation is fullled. On this basis, a tax register will be established for the village and each individual will be taxed accordingly. In a declaration of death, the family of the deceased must request action from the village scribe: as long as this is not done, the family answers for the payment of the dead person’s taxes; and after the scribe has ascertained the correctness of the declaration, he will remove the name from the list of persons liable for taxation. E.g. P.Petaus 7.15–17 (= C.Pap.Gr. 62; Ptolemais Hermou, A.D. 185): διὸἀξιῶ ταγῆναι αὐτοῦ τὸὄνο(μα) ἐντῇ τῶν τετελευτηκ(ότων) τάξει ὡς ἐπὶ τῶν ὁμ{ι}οίων ‘Therefore I request that his name be placed in the list of deceased persons, following the usual practice’. This is much closer to what we observe in the certicates of sacrice, where individuals expect some action on the part of the commissioners, namely that they sign the certicate. E.g. SB I

73 See Clarke 1984: 33. 74 2 See e.g. the case of P.Gen. I 16.18–21 (Soknopaiou Nesos, A.D. 207) a petition that refers to an imperial edict enforced by Subatianus Aquila, Prefect of Egypt (in charge between A.D. 206 and 211): τοῦ λαμπροτ̣ά̣του ἡγεμόνος ΣουβατιανοῦἈκύλα κελεύσαντος πάντας τοὺς ἀπὸ ξένης ὄντας κατ̣ισελθεῖνεἰςτὴν ἰδίαν ἐχομένους τῶν συνηθῶν ἔργων ‘Subatianus Aquila, uir clarissimus, ordered that all those living away (from their villages) should go back to their places of origin and tend to their customary works’. 75 On the format of declarations of death, see C.Pap.Gr. = Casarico 1985:10–21. Many more declarations of death were published in the past thirty years. 76 See Rives 1999: 150, n. 82.

TABLE 7

CENSUS RETURN DECLARATION OF DEATH CERTIFICATE OF SACRIFICE Heading: addressee Heading: addressee Heading: addressee (authority) + (authority) + applicant (authority) + applicant applicant Declaration: house + Declaration: death of one or Declaration: description of past household (persons) several individuals behaviour and present action No request; the applicant Request that the person(s) be Request for the commissioners’ merely submits the removed from the register signature document Control by a village scribe The commissioners’ assistant introduces the signature in their name No signature Signature Signature Date Date Date

4440.13–14 (15; Theadelphia, A.D. 250): ἀξιῶὑμᾶς ὑποσημιώσασθαί μοι ‘I request that you sign for me’. In spite of the distinct purpose of all three types of document, the certicates of sacrice remain a kind of declaration. This stands out in particular in the libelli from the Oxyrhynchite nome: in one case, the word ἀπογραφή is explicitly used;77 and two libelli use the wording ἐπιδέδωκα ‘I have submitted’.78 There is little doubt that the commission kept a copy of every declaration. The applicant, however, was presumably handed a signed copy, which he would have had to produce on request.

VI OVERALL CONCLUSION

The purpose of the certicate of sacrice was different from a census declaration. It was conceived as a way to ensure not mere declaration, but actual participation. On the one hand the declarative part, which related to past conduct, bore a rather formal character. On the other hand, the performative dimension was essential in a libellus: the procedure was devised in such a way that it was difcult for villagers to escape any active involvement. Passive partaking was not enough, since everyone had to ingest the offering and then ask for a written conrmation from the commissioners. The special layout of the ὑπόμνημα, with its window left blank for signature, nds no exact parallel in the thousands of other preserved papyri from Egypt.79 Villagers were thus forced to secure help from a professional scribe in order to prepare their own act of submission to the ritual, which was duly observed by representatives of the state authorities. At the level of the Empire, Brent was probably right to describe Decius’ edict as a collective supplicatio aimed at regaining the favour of the gods. At the level of an Egyptian village, this mattered little. The local administration had to nd an effective mechanism by which every villager would comply with the order. Decius, in his brief reign, lacked the necessary time to rally the Empire, but this failure cannot be attributed

77 See 44 = P.Oxy. LVIII 3929, verso: ἀπογρ(αφὴ) Ἀμοϊτᾶ μητ(ρὸς) Τααμόϊτ(ος). 78 See 8 = P.Oxy. XII 1464.15; 44 = P.Oxy. LVIII 3929.20. 79 The receipts for dyke tax — also called ‘penthemeros-certicates’—cannot be compared with the certicates of sacrice: although they were prepared in advance, to have personal details lled in later, their structure is much simpler and they do not involve the relatively complex action that is described in the certicates of sacrice. On penthemeros-certicates, see Sijpesteijn 1963 and Kruse 2002: 306–19.

S to ’ PPLICANT Continued GENDER A AY D ONTH M OME N ROVENANCE P Theadelphia Ars. Payni 18 m ]/ 8 RINT [ ABLE P ( T 2 .) NO ]/ IGITAL 3 [ MAGE AVAILABLE I D 125 p Ptolemais Euergetis Ars. no date no date f Chr. EFERENCE R P.Lips. II 152SB VI 9084 p no Euhemeria Euhemeria Ars. Ars. Payni 22 Payni 23 m f EYER M NIPFING 387 168 ?9 P.Hamb. 2 I 61a11 326 SB no 4 I 4436 635 SB I 4437 6a12 SB I 4438 SB I13 4440 p SB I 5943 714 no 8 P.Mich. III no 157 SB 9 I p 4441 Theadelphia SB p I d 4442 SB I 4443 Ars. no Theadelphia no Payni Theadelphia no 19 Theadelphia Ars. Theadelphia Ars. Theadelphia Theadelphia Payni Ars. Payni 20 Ars. Theadelphia 20 Payni Ars. Ars. Theadelphia 21 Payni f Payni Theadelphia Payni 22 Ars. 22 23 Ars. Payni f Ars. 23 Payni f m 25 Payni 25 K 125323W. 21 25 BGU I 287 P.Ryl. I 12 p p d Alexandrou Nesos Ptolemais Euergetis Ars. Ars. Epeiph Payni 2 20 m f 374332 246 P.Wisc. II 87 P.Oxy. IV 658 22 P.Oxy. XII 1 1464 d SB no I 4455 no SB I 4435 d d Oxyrhynchus Narmouthis Oxyrhynchus Oxy. Ars. Oxy. Payni Philadelphia Payni Epeiph 20 3 10 Ars. m m m no date no date m 1 UMBER N 11 12 13 14 15 16 17 18 19 20 1 2 3 4 5 6 7 8 9 10

http:/www.cambridge.org/core Table 8 Continued NTEFR N OTN FTECRIIAE FPGNSACRIFICE PAGAN OF CERTIFICATES THE OF CONTENT AND FORM THE ON 1 NUMBER KNIPFING MEYER REFERENCE IMAGE AVAILABLE(P[RINT]/ PROVENANCE NOME MONTH DAY APPLICANT’S 2 D[IGITAL]/NO.) GENDER 21 27 P.Ryl. I 112a d Theadelphia Ars. Payni 26 f . http://dx.doi.org/10.1017/S0075435816000617 22 36 P.Mich. III 158 d Theadelphia Ars. Payni 27 f 23 15 10 SB I 4444 p Theadelphia Ars. Payni 27 m . University of Geneva 24 39 P.Hamb. I 61b no Theadelphia Ars. Payni 27 25 40 SB III 6827 d Theadelphia Ars. Payni 21 26 29 P.Ryl. II 112c d Theadelphia Ars. Payni 28 f 27 5 11 SB I 4445 p Theadelphia Ars. Payni 28 f , on 28 21 Nov2016 at12:36:55 16 12 SB I 4446 no Theadelphia Ars. Payni 29 29 17 13 SB I 4447 no Theadelphia Ars. Payni 29 30 18 14 SB I 4448 no Theadelphia Ars. Payni 29 m 31 19 15 SB I 4449 no Theadelphia Ars. Payni 29 32 34 PSI V 453 p Theadelphia Ars. Payni 2- f , subjectto theCambridgeCore termsofuse,available at 33 10 5 SB I 4439 p Theadelphia Ars. Payni no date m 34 30 15a P.Meyer 15 p Theadelphia Ars. Epeiph 3 f 35 20 16 SB I 4450 p Theadelphia Ars. Epeiph 20 m 36 28 P.Ryl. II 112b d Theadelphia Ars. no date no date m 37 24 20 SB I 4454 no Theadelphia Ars. no date no date 38 32 20b P.Meyer 17 no Theadelphia Ars. no date no date f 39 21 17 SB I 4451 no Theadelphia Ars. no date no date f 40 31 20a P.Meyer 16 no Theadelphia Ars. no date no date f 193 41 23 19 SB I 4453 no Theadelphia Ars. no date no date Continued http:/www.cambridge.org/core/terms Downloaded from 194 Table 8 Continued 1

http:/www.cambridge.org/core NUMBER KNIPFING MEYER REFERENCE IMAGE AVAILABLE(P[RINT]/ PROVENANCE NOME MONTH DAY APPLICANT’S 2 D[IGITAL]/NO.) GENDER 42 41 SB III 6828 d Theadelphia Ars. no date no date 43 22 18 SB I 4452 no Theoxenis Ars. no date no date f . http://dx.doi.org/10.1017/S0075435816000617 44 P.Oxy. LVIII 3929 d Thosbis Oxy. Epeiph no date m 45 PSI VII 778 d ? Ars. (?) Epeiph 2 f

. 46 University of Geneva P.Oxy. XLI 2990 d Oxyrhynchus ? Oxy. no date no date 1 The numbering follows the order provided by Reinhold Scholl in P.Lips. II 152. 2 Status as of April 2016. The reference to most images, both printed and digital, was recorded on the website . Digital images of papyri kept in the Rylands collection (Manchester) did not appear there, but could be found on the website of the University of Manchester Library Image Collections . SCHUBERT PAUL 3 10 , on An image of was found fortuitously on the temporary webpage of an exhibition at the Staats- und Universitätsbibliothek in Hamburg. 21

21 Nov2016 at12:36:55 Although of low quality, it displayed a shape indicating that this papyrus had been found squeezed together with . , subjectto theCambridgeCore termsofuse,available at ON THE FORM AND CONTENT OF THE CERTIFICATES OF PAGAN SACRIFICE 195

FIG. 1. 1 = BGU I 287 (Alexandrou Nesos, Arsinoite nome). (© Staatliche Museen zu Berlin – Ägyptisches Museum und Papyrussammlung, P 7297: with permission)

FIG. 2. 16 = SB I 5943 (Theadelphia, Arsinoite nome). (© Staatliche Museen zu Berlin – Ägyptisches Museum und Papyrussammlung, P 13430, photograph taken by Sandra Steiß: with permission)

FIG. 3. 44 = P.Oxy. LVIII 3929 (Thosbis, Oxyrhynchite nome). (© Courtesy of the Egypt Exploration Society and Imaging Papyri Project, Oxford)

the lower levels of his administration, where every element in the procedure testies to the effectiveness of the system.

University of Geneva [email protected]

BIBLIOGRAPHY Bagnall, R. S., and Frier, B. W. 1994: The Demography of Roman Egypt, Cambridge Barnes, T. D. 2007: ‘Constantine after seventeen hundred years: the Cambridge Companion, the York exhibition and a recent biography’, International Journal of the Classical Tradition 14, 185–220 Brent, A. 2010: Cyprian and Roman Carthage, Cambridge Casarico, L. 1985: Il controllo dell popolazione nell’ Egitto romano. 1. Le denunce di morte, Azzate

Clarke, G. W. 1969: ‘Some observations on the persecution of Decius’, Antichthon 3, 63–78 (some passages of this article were taken over in Clarke 1984) Clarke, G. W. 1984: The Letters of St. Cyprian of Carthage, Vol. I, New York Corcoran, S. 2015: ‘From unholy madness to right-mindedness: or how to legislate for religious conformity from Decius to Justinian’, in A. Papaconstantinou with N. McLynn and D. L. Schwartz (eds), Conversion in Late Antiquity: Christianity, Islam, and Beyond: Papers from the Andrew W. Mellon Foundation Sawyer Seminar, University of Oxford, 2009–2010,Farnham/Burlington, 67–94 Foley, H. 1999: The Homeric Hymn to Demeter, Princeton Fournet, J.-L. 2012: ‘Homère et les papyrus non littéraires: le poète dans le contexte de ses lecteurs’, in G. Bastianini and A. Casanova (eds), I Papiri omerici. Atti del convegno internazionale di studi, Firenze, 9–10 giugno 2011, Florence, 125–57 Geoltrain, P. 2005: ‘Roman pseudo-clémentin. Introduction’, in P. Geoltrain and J.-D. Kaestli (eds), Écrits apocryphes chrétiens II, Paris, 1175–87 Keresztes, P. 1975: ‘The Decian libelli and contemporary literature’, Latomus 34, 761–81 Kienast, D. 1996: Römische Kaisertabelle. Grundzüge einer römischen Kaiserchronologie, Darmstadt Knipng, J. R. 1923: ‘The libelli of the Decian persecution’, Harvard Theological Review 16, 345–90 Kruse, Th. 2002: Der königliche Schreiber und die Gauverwaltung. Untersuchungen zur Verwaltungsgeschichte Ägyptens in der Zeit von Augustus bis Philippus Arabs (30 v. Chr.–245 n. Chr.), Archiv für Papyrusforschung Beiheft 11, Leipzig Lane Fox, R. 1986: Pagans and Christians, Harmondsworth Leadbetter, W. L. 1982: ‘A libellus of the Decian persecution’, in G. H. R. Horsley (ed.), New Documents Illustrating Early Christianity, 1977, North Ride, Macquarie University, 180–5 Legras, H. 1907: La Table latine d’Héraclée, Paris Lenski, N. 2006: ‘Introduction’, in N. Lenski (ed.), The Cambridge Companion to the Age of Constantine, Cambridge, 1–13 Luijendijk, A. M. 2008: Greetings in the Lord: Early Christians and the Oxyrhynchus Papyri, Cambridge, MA Meyer, P. M. 1910: Die libelli aus der decianischen Christenverfolgung, Abhandlungen der Preussischen Akademie der Wissenschaften 3, Anh. 5, Berlin Molthagen, J. 1975: Der römische Staat und die Christen im zweiten und dritten Jahrhundert, Göttingen Popkes, W. 1976: ‘Gemeinschaft’, Reallexikon für Antike und Christentum 9, 1116–17 Primavesi, O. 1996: ‘Zur Geschichte des deutschen Papyruskartells’, Zeitschrift für Papyrologie und Epigraphik 114, 173–87 Rathbone, D. 1991: Economic Rationalism and Rural Society in Third-Century A.D. Egypt, Cambridge Rathbone, D. 2009: ‘The rst acquisition: the archive of Heroninos’, in G. Bastianini and A. Casanova (eds), 100 anni di istituzioni orentine per la papirologia, Florence, 17–29 Richardson, N. J. 1974: The Homeric Hymn to Demeter, Oxford Rives, J. B. 1999: ‘The decree of Decius and the religion of empire’, Journal of Roman Studies 89, 135–54 Rubensohn, O. 1902: ‘Griechisch-römische Funde in Ägypten’, Archäologischer Anzeiger,46–9 Schubert, P. 2007: Philadelphie. Un village en mutation entre le IIe et le IIIe s. ap. J.-C., Basle Schwartz, J. 1947: ‘Une déclaration de sacrice du temps de Dèce’, Revue Biblique 54, 365–9 Selinger, R. 1994: Die Religionspolitik des Kaisers Decius. Anatomie einer Christenverfolgung, Frankfurt a.M./Berlin/Bern/New York/Paris/Vienna Sijpesteijn, P. J. 1963: Penthemeros-Certicates in Graeco-Roman Egypt, Papyrologica Lugduno-Batava 12, Leiden Youtie, H. C. 1971: ‘“Bradeos graphon”: between literacy and illiteracy’, Greek, Roman and Byzantine Studies 12, 239–61 (repr. in H. C. Youtie, Scriptiunculae II.30 (Amsterdam, 1973), 629–51)